“Gas filter correlation radiometry” (GFCR) is an optical remote sensing method used to produce highly sensitive measurements of “targeted” gases. A conventional GFCR measurement system using two single-element detectors is shown in FIG. 1 and is referenced generally by numeral 10. The basic elements of GFCR system 10 form an optical train that includes:
focusing optics (e.g., a telescope) 12 that focuses the image contained in light 200 onto a field stop 14 that sets the field-of-view of the light focused by optics 12,
optics 16 that collimate the light passing through field stop 14,
a chopper 18 used to modulate the collimated light,
a spectral filter 20 that confines the light (collimated by optics 16) to a specific spectral bandpass where (spectrally) a gas of interest absorbs,
optics 22 for splitting the spectrally filtered light into two paths 24 and 26 where path 24 defines a region that is non-absorbing within the spectral bandpass at which the gas of interest absorbs,
a single-element, light-intensity detector 28 disposed in path 24, and
a gas cell 30 filled with a gas of interest (i.e., the target gas) and disposed in path 26 such that the light traveling therealong passes through gas cell 30 prior to impinging on a single-element, light-intensity detector 32.
GFCR system 10 further uses “back-end” electrical components that includes balancing electronics 40 coupled to the outputs of detectors 28 and 32. In general, balancing electronics 40 include a balancing amplifier 42, a differential amplifier 44 and a gain amplifier 46 that cooperate to measure a difference between the outputs of detectors 28 and 32.
GFCR system 10 uses a sample of the gas to be detected (i.e., the target gas) as a filter for removing sensitivity to that gas in path 26. That is, the light is passed through field stop 14, collimated, and spectrally filtered with broadband filter 20 to confine the light to a spectral bandpass where the target gas absorbs. After the beam is split by optics 22, gas cell 30 absorbs light from spectral wavelengths coinciding with spectral absorption features (i.e., typically spectral absorption lines) of the target gas.
In practice, the detector signals from detectors 28 and 32 are electronically balanced to be approximately equal when viewing light 200 from an unattenuated light source such as the sun observed above the atmosphere from a satellite. Then, when the light source is observed through the atmosphere during solar occultation, a difference between the two signals is induced and measured. Absorption by the target gas in the observed scene attenuates the vacuum path signal generated by detector 28. However, the gas path signal generated by detector 32 is minimally attenuated. The difference signal (i.e., the difference between the two signals generated by detectors 28 and 32) is highly sensitive to and correlated with the amount of target gas in the line-of-sight of GFCR system 10.
Useful GFCR measurements must be tailored to the absorption characteristics of the target gas, and depend on the ability to maintain a stable and calibrated gas cell containing a sample of the target gas. For example, when the GFCR method was employed in a solar occultation experiment (i.e., the “halogen occultation experiment” or HALOE), sensitivities of 10−5 in mean band absorption were achieved by a system similar to that depicted in FIG. 1. The two detectors' signals were differenced and balanced by electronics 40 to give nearly zero difference during solar observation above the atmosphere. The difference signals measured during solar occultation were used very successfully as measures of target gas absorption. To achieve high precision, the difference signals included an additional gain of one hundred or more. The key to making these measurements is the ability to determine the balance and rate of change of the difference signal immediately before the observation, which mitigates error due to drifts in detector response. To achieve the desired measurement accuracy of 1 part in 105, the balance must be known to 10−5 of the full broadband signal. Thus, small drifts in detector response, if not detected and corrected, can severely corrupt the difference measurement.
In addition to use in solar occultation, there has been hope that GFCR could be used for solar backscatter measurements, which could yield much better geographical coverage than solar occultation and be realized using small commercial devices. However, because the conventional two-detector method requires continuous high-precision calibration of the balance condition (i.e. calibration of the signal drift due to changes in system response), most researchers have abandoned the two-detector method in favor of single-detector methods. Unfortunately, while single-detector methods can nearly eliminate detector instability as an error source by measuring both signals with the same detector, they introduce a host of other problems, depending on method of implementation. For example, if the gas cell condition is modulated by changing pressure or optical mass, there is a significant decrease in sensitivity because the cell modulation produces a relatively small spectral difference between paths. The signals are also difficult to model because of gas heating and cell state variation that may not reach uniform equilibrium.
In another single-detector method, the light path is switched between a gas-cell path and a non-gas-cell (e.g., vacuum) path by either rapidly re-routing the beam (e.g., polarization switching techniques) or moving the gas cell into and out of the beam. However, both of these approaches introduce noise due to beam steering and loss of signal integration time due to time between modulated states.
An even greater problem with any single-detector method is the loss of measurement simultaneity and/or the ability to exactly match field-of-views for gas and vacuum paths. If the scene changes during the time necessary to switch between modulated states or because of field-of-view mismatch, the change in normalized difference signal (caused by scene brightness variation) will corrupt the data interpretation that assumes the difference signal is produced solely by spectral variation. For example, a satellite traveling at 7 km/sec encountering a 1% per kilometer change in mean scattering brightness over the field-of-view will experience a fractional brightness change of 10−4 in 1.4 milliseconds which would be falsely interpreted as spectral variation. This presents a severe problem for the single-detector method, or any method that does not make simultaneous and spatially identical measurements of the two states (i.e., gas path and vacuum path).